Skip to main content

Table 5 Simulation results evaluating Rule 3 when the threshold probability is set at 0.90

From: Evaluation of a multi-arm multi-stage Bayesian design for phase II drug selection trials – an example in hemato-oncology

Sample size True benefit Posterior mean estimate biases Average sample sizes % Early stopping
  dB dC pA pB pC A B C B C
40 −0.15 0.00 0.0092 0.0182 0.0254 39.9621 39.8136 37.9895 0.54 % 6.45 %
  −0.05 0.00 0.0088 0.0237 0.0272 39.8054 38.8511 37.8795 3.56 % 6.90 %
  0.00 0.00 0.0085 0.0254 0.0270 39.6095 37.9435 37.9783 6.69 % 6.53 %
  0.05 0.00 0.0089 0.0324 0.0250 39.5103 36.5882 38.0288 11.66 % 6.37 %
  0.10 0.00 0.0062 0.0384 0.0264 39.2740 34.3440 38.0311 19.85 % 6.37 %
  0.15 0.00 0.0056 0.0416 0.0254 39.0452 31.9203 38.0610 29.96 % 6.22 %
  0.20 0.00 0.0049 0.0451 0.0256 38.8153 28.6999 37.9511 42.93 % 6.61 %
  0.25 0.00 0.0051 0.0423 0.0254 38.5798 25.0534 37.9815 57.41 % 6.60 %
  0.30 0.00 0.0052 0.0369 0.0250 38.4489 21.2180 38.0583 71.85 % 6.33 %
  0.35 0.00 0.0037 0.0261 0.0270 38.2338 17.3342 37.9999 83.87 % 6.49 %
  0.40 0.00 0.0039 0.0098 0.0265 38.0706 14.1594 37.8966 92.38 % 6.79 %
  0.45 0.00 0.0035 −0.0115 0.0266 38.0738 11.1299 37.9628 97.64 % 6.54 %
100 −0.15 0.00 0.0029 0.0093 0.0226 99.8389 99.4843 94.1617 0.55 % 6.49 %
  −0.05 0.00 0.0031 0.0169 0.0237 99.3256 96.7377 93.7367 3.55 % 6.94 %
  0.00 0.00 0.0020 0.0233 0.0222 99.0139 93.7839 94.1902 6.91 % 6.48 %
  0.05 0.00 0.0013 0.0322 0.0224 98.3338 89.3276 94.1794 12.36 % 6.48 %
  0.10 0.00 0.0012 0.0405 0.0227 97.6239 82.2664 94.1533 21.90 % 6.50 %
  0.15 0.00 −0.0009 0.0509 0.0234 96.5788 71.2643 94.0376 37.94 % 6.60 %
  0.20 0.00 −0.0025 0.0578 0.0235 95.7497 56.8519 94.0065 59.48 % 6.66 %
  0.25 0.00 −0.0025 0.0577 0.0230 95.1262 41.9689 93.9749 79.99 % 6.71 %
  0.30 0.00 −0.0030 0.0473 0.0235 94.5634 29.7273 93.9678 92.99 % 6.71 %
  0.40 0.00 −0.0025 0.0119 0.0239 94.0066 15.1126 93.8206 99.84 % 6.79 %
  0.45 0.00 −0.0027 −0.0110 0.0227 94.2995 11.7989 94.1851 100.00 % 6.48 %
  1. pA, pB and pC refer to the posterior means of response probability in arms A, B and C, respectively; δ*=0.15. Bold data refer to the null hypothesis of absence of any treatment difference (dB=dC=0), or to the alternative hypothesis of an expected true 0.15 benefit of treatment B (dB=0.15)